Query construction #91
Conversation
Possible to construct term/boost/boolean queries
pyserini/search/pyquerybuilder.py
Outdated
|
|
||
|
|
||
| def get_boolean_query_builder(): | ||
| """ Get a BooleanQueryBuilder object. |
There was a problem hiding this comment.
extra space here, per PEP 8. and below.
pyserini/search/pyquerybuilder.py
Outdated
|
|
||
|
|
||
| def get_clause_should(): | ||
| """ Get a BooleanClause.Occur.SHOULD statement |
There was a problem hiding this comment.
PEP8 - docstring comments should end in periods.
pyserini/search/pysearch.py
Outdated
| @@ -72,14 +72,14 @@ class SimpleSearcher: | |||
| def __init__(self, index_dir: str): | |||
| self.object = JSimpleSearcher(JString(index_dir)) | |||
|
|
|||
| def search(self, q: str, k: int=10, t: int=-1, | |||
| def search(self, q, k: int=10, t: int=-1, | |||
There was a problem hiding this comment.
We can still provide a type hint here? using Union?
pyserini/search/pyquerybuilder.py
Outdated
| return JQueryGeneratorUtils.getBooleanClauseShould() | ||
|
|
||
|
|
||
| def get_clause_must(): |
There was a problem hiding this comment.
Would it be better to implement an Enum for these? Like here? https://github.com/castorini/pyserini/blob/master/pyserini/pyclass.py#L103
There was a problem hiding this comment.
Hrm. Wait, or maybe somehow bind them to singleton symbols?
Dunno...
tests/test_querybuilding.py
Outdated
| hits1 = self.searcher.search(bq1) | ||
| hits2 = self.searcher.search('information retrieval') | ||
|
|
||
| self.assertEqual(hits1[0].docid, hits2[0].docid) |
There was a problem hiding this comment.
Can you iterate through all the hits to check all scores and docids?
| self.assertEqual(hits1[0].docid, hits2[0].docid) | ||
| self.assertEqual(hits1[0].score, hits2[0].score) | ||
|
|
||
| boost3 = pyquerybuilder.get_boost_query(term_query1, 2.) |
There was a problem hiding this comment.
Maybe add a test case where you assign both weight 2 - docid (in entire hits) should be the same, but scores different?
Make it possible to construct queries in Pyserini